generated: '2026-07-19'
method: searched
source: https://api.exactly.ai/public/docs/ + openapi/exactly-ai-public-openapi-original.json
authentication:
  style: http bearer token
  scheme: APITokenAuth
  header: 'Authorization: Bearer <token>'
  availability: enterprise clients
  ref: authentication/exactly-ai-authentication.yml
versioning:
  scheme: uri-path
  current: v1
  example: /public/v1/models/
pagination:
  style: limit-offset
  params:
  - name: limit
    default: 100
    min: 1
    max: 5000
  - name: offset
    default: 0
    min: 0
  response_fields:
  - items
  - count
  - total_count
  schema: PaginatedResponse
async_jobs:
  pattern: submit-then-poll
  description: Long-running operations (model training, image upscale, vectorize, remove-bg) return 202
    Accepted and are polled via a corresponding GET endpoint until the result is ready (200). Training
    exposes a dedicated progress endpoint.
  poll_endpoints:
  - GET /public/v1/models/{uid}/train/progress/
  - GET /public/v1/images/{uid}/upscales/{scale}/
  - GET /public/v1/images/{uid}/vectors/
  - GET /public/v1/images/{uid}/remove-bg/
identifiers:
  type: uuid
  fields:
  - uid
  - model_uid
  - image_uid
idempotency:
  supported: false
  note: No idempotency-key header or parameter is documented in the Public API.
error_envelope:
  shape: DurerError {message, client_message}
  ref: errors/exactly-ai-problem-types.yml
rate_limit_signaling:
  documented: false
cross_links:
  errors: errors/exactly-ai-problem-types.yml
  lifecycle: lifecycle/exactly-ai-lifecycle.yml
  authentication: authentication/exactly-ai-authentication.yml